azonenberg changed the topic of #homecmos to: Homebrew CMOS and MEMS foundry design | Wiki: http://homecmos.drawersteak.com/wiki/Main_Page | Repository: http://code.google.com/p/homecmos/ | Logs: http://en.qi-hardware.com/homecmos-logs/
jamie has quit [Excess Flood]
jamie has joined #homecmos
jamie has quit [Excess Flood]
jamie has joined #homecmos
mrdata has joined #homecmos
Textmode has quit [Quit: "It was one dev, naked in a room with a carton of cigarettes, a thermos full of coffee and bourbon, and all his summoned angels."]
chris_99 has joined #homecmos
jamie has quit [Excess Flood]
jamie has joined #homecmos
<_Sync_> whitequark: did you read housekeepers work on ceramic to metal seals and the like=?
<whitequark> "Housekeeper seal" ?
<whitequark> so... metal-glass-ceramic
chris_99 has quit [Remote host closed the connection]
<_Sync_> he has a few books on that
<whitequark> mm i see
<whitequark> I am not sure I see how it would be easier though, kovar seals are like $40 each and they are fairly nontrivial to make yourself
<_Sync_> well he also talks in depth about all kinds of glass or ceramic work
<_Sync_> in the end the key to make good ceramic/metal seals is a hydrogen furnance
<whitequark> mm yes i should read that book
<whitequark> hydrogen is... annoying to work with
<_Sync_> not really, if you have a proper oven it is not really all that dangerous to work with
<whitequark> I mean more like having a cylinder of it
<_Sync_> just making a safe oven is annoying
<_Sync_> why?
<whitequark> if it leaks, it can get rather nasty
<whitequark> hot invisible flame, not to mention the enormous explosive mixture range
<_Sync_> sure, but why would it leak
<_Sync_> modern gas bottles are very safe
<_Sync_> not sure how modern they are in russia tho ;)
<whitequark> it depends but i believe they're on the order of ones used in EU
<whitequark> unless you buy really old shit
<whitequark> i guess i could just synthesize it on spot, in the tiny quantities needed for the vacuum oven
<_Sync_> the quantities are not all that small
<_Sync_> you need to purge through the entire heat cycle
<_Sync_> with rates of maybe 5-20l/min
<whitequark> hm i see
<_Sync_> and fairly high purities
jamie has quit [Excess Flood]
jamie has joined #homecmos
mrdata_ has joined #homecmos
mrdata has quit [Ping timeout: 264 seconds]
<SpeedEvil> I wondered a while back about InGa with nanoscopic ferrite particles in, as vacuum-tight constrainable by magnetic field 'zero' leak moving seals
<SpeedEvil> ^ferrofluid
<whitequark> aren't ferrofluid passthroughs a thing?
<whitequark> like, already
<SpeedEvil> yes
<_Sync_> yup
<_Sync_> they are actually pretty spendy :X
<SpeedEvil> Exactly
<SpeedEvil> It's a pity printable magnetics suck a bit.
<SpeedEvil> But at least magnetic steel is easy to machine
<whitequark> is there printable anything that doesn't suck?
<whitequark> PLA sucks. SLA is... okay-ish quality, slow, and the resin puts printer ink to shame. printing brass or steel results in something disgustingly porous
<whitequark> selective electron beam sintering in vacuum seems to be okay, but incredibly expensive
<SpeedEvil> whitequark: Inconel
<SpeedEvil> and yeah - I've been wondering on that too
<SpeedEvil> A ~2A or so electron gun doesn't seem _horribly_ complex.
<whitequark> you need a turbo pump
<whitequark> well, at EHSM there was a guy from Warsaw's ITR which did it in a sort of roundabout way
<whitequark> he used the residual gas to do /ion/ beam sintering
<whitequark> so it worked with only a rotary vane pump
<SpeedEvil> sublimation pumps can get down to quite reasonably low pressures in some cases.
<SpeedEvil> If you can get on top of leakage and your print media doesn't outgas under load
<SpeedEvil> ^heat
<whitequark> hrm
<whitequark> print media very certainly does, by virtue of being a fine powder with a shit ton of surface area
<whitequark> don't forget you need to add print media too...
<SpeedEvil> I don't know how much gas inconel or SS or ... beads will sorb
<whitequark> 'beads' ?
<SpeedEvil> for sintering
<whitequark> well.
<whitequark> let's say 1 thou resolution
<whitequark> so 0.3 thou particle size
<SpeedEvil> I'm not sure you can do that
<whitequark> why not?
<whitequark> you can focus the beam pretty well
<SpeedEvil> Well, for one thing, 1 thou layers is going to take for fucking ever.
<whitequark> layers would be thicker.
<SpeedEvil> yes, but they can't be that much thicker
<whitequark> you can actually adjust resolution pretty easily
<SpeedEvil> Also, 'conduction' through radiation is quite efficient at thosse scales.
<whitequark> since beam power and spot size are both trivially controlled
<_Sync_> SpeedEvil: you will need some kind of active pump
<SpeedEvil> As particle size drops, square/cube means that it is a much more efficient radiator
<_Sync_> sorption pumps have very bad absolute capacities
<_Sync_> and turbos are cheap
<whitequark> "cheap"
<_Sync_> they really are
<whitequark> not if you're trying to mass-produce 3D printers
<_Sync_> did not pay more than 150€ for any of mine
<SpeedEvil> _Sync_: yes - I don't have a good handle on gas loads from the media. And capacity varies by gas. Backstreaming through the pump is a concern that I was wondering about addressing the very silly way of enclosing the oil pump and putting it in a nitrogen box.
jamie has quit [Excess Flood]
<whitequark> _Sync_: they were used.
<SpeedEvil> But that's likely overkill.
<SpeedEvil> And that too
<_Sync_> sure, but there is not much to go wrong besides the bearings
<_Sync_> and even new
<whitequark> the secondary market will not support anything in batches
jamie has joined #homecmos
<_Sync_> a 150l/s pump runs you around 4k new
<whitequark> yeah
<_Sync_> that's pretty cheap after all
<_Sync_> SpeedEvil: use a dry scroll then
<whitequark> it's not a mass market device even for a vrey conservative notion of "mass"
<_Sync_> but metal powders have a LOT of trapped gas
<_Sync_> because they have a high surface area
<SpeedEvil> _Sync_: I have a nice edwards double stage oil.
<whitequark> and if not, I mean, those metal sintering in vacuum printers are already made and sold
<_Sync_> thus they have a lot of OH groups on the wurface
<whitequark> SpeedEvil: not enough for EB I believe
<SpeedEvil> _Sync_: I know - but I don't know if they do once you heat it red hot for an hour
<_Sync_> well, after outgassing they get better
<_Sync_> but everytime you vent you have to bake again
<_Sync_> I switched all my pumping systems over to dryscroll/turbo
<SpeedEvil> Unless you keep your media under vacuum or dry gas, which isn't quite prohibitive.
<_Sync_> because I don't want to deal with the oil vapors anymore
<whitequark> SpeedEvil: argon will still be trapped on the surface
<SpeedEvil> ^Above opinions were formed when I was investigating relatively high presure uses like making mirrors. I don't have a good handle on UHV
<whitequark> at least a monolayer
<SpeedEvil> whitequark: I was assuming you could pick an easily bakeable gas - I don't off-hand know what it'd be
<whitequark> if you have something like 1m^3/g
<whitequark> er
<_Sync_> argon is not too bad
<whitequark> m^2
<whitequark> so like 8m^2/ml
<whitequark> that is going to be a lot
<whitequark> also argon will not be sequestered by sublimation pumps or sth, you need a kinetic one
<_Sync_> well, sublimators will still pump noble gasses
<_Sync_> although they like reactive ones better
<SpeedEvil> whitequark: yes, which is why I said N2 initially
<SpeedEvil> ^thought
<SpeedEvil> Speed kind-of-sucks quite badly with nobles.
jamie has quit [Excess Flood]
jamie has joined #homecmos
<whitequark> N2 will form nitrides though
<whitequark> rather well at that
<whitequark> which will make your finished piece extremely brittle
<SpeedEvil> I was assuming it would bake out
<SpeedEvil> I've not actually researched this properlym and would ned to do so before seriously...
<whitequark> bakeout could work
<whitequark> however apparently you do not need a lot of N2 to get a shitty crystal
<whitequark> like there's a reason tig welding requires pretty pure argon
<SpeedEvil> I would say hydrogen - but... :)
<whitequark> uh
<whitequark> hydrogen embrittlement is also a thing
<SpeedEvil> yeah
<SpeedEvil> hence the ...
Textmode has joined #homecmos
<whitequark> huh, before 1952, it was not understood that metal-ceramic seal is not merely mechanical
<_Sync_> 0o
<whitequark> exactly
<whitequark> "In fact, at the time this
<whitequark> picture was presented, attention was called to the absence of
<whitequark> reaction between the metalizing and the ceramic, leading to
<whitequark> the general agreement that the bond in seals was purely mechanical
<whitequark> gripping. "
<whitequark> well, maybe not 1952, more like late 40s
<whitequark> that's not all
<whitequark> until mid-50s metallography was not understood as a useful method for examining ceramics!
<_Sync_> hm, understandable